This study aimed at evaluating the supervised pixel-based classification of the maximum likelihood in the examination of the climactic changes of the Urmia Lake catchment area in 1-, 6-, 10-, and 14-year spans and then evaluating and modeling the changes in the gardens neighboring Lake Urmia during a 30-year span and its role in the changes at Lake Urmia water level. In this study, first the Landsat images of the years 1990, 2000, 2014, 2020, and 2021 were downloaded. Then, using the ENVI5.3 software, a classification was made based on the maximum likelihood method. Next, the IDRISSI TERRSET software and CA-MARKOV model were used to model conditions for the year 2051. This model was then analyzed in the GIS software. It was found that the classification based on the maximum likelihood method has been an appropriate one. The results of this classification showed that overall, the maximum changes from 1990 to 2021 has been related to the irrigated gardens and fields. In this period, 3495 square kilometers have been added to these lands, i.e., the number has doubled. Modeling the 2051-year conditions in the light of the transition probability matrix showed that the maximum likelihood of land use changes is in the irrigated garden and farm use. On the other hand, the modeling in this study indicated that there has been a reduction in the expanse of irrigated gardens and fields as well as salt lands in the region. Finally, the model predicted an increase in water area in 2051.